documented, excela has a program and we have a healthy -- the cloud is delivering good performance as we have tested it. we have a thing we recognized and brian talked about it and i want to share, the excela team has rebuilt the original code. when we restarted this project there was a very important assumption made when the original time line etc. was determined. they started to build on that testing it and implementing it we found significant problems and one was the fikac lator, as brian eluded to, it was clumsy, slow and included a lot of custom configuration which would be a maintenance problem into the future. so the developers rewrote it. and they rewrote it and it's much faster, the d.b.i. technical team thinks they will do a good job maintaining it. this is incredibly important. the other challenge we faced was with the address database. i had a concern about this myself and i saw the architecture of it. it created a point of failure, the way the system was designed. it also could create latency between when we start a permit and have to pick up an address so i was already c